Impact of N-acetyl-l-cysteine on SARS-CoV-2 pneumonia and its sequelae: results from a large cohort study

To the Editor:

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection may cause pneumonia and acute respiratory distress syndrome (ARDS), whose pathogenesis has been partially related to an increased systemic inflammatory response with great production of pro-inflammatory cytokines causing a “cytokine storm” and an oxidative stress imbalance [1].

N-acetyl-l-cysteine (NAC) is a precursor of reduced glutathione [2] that has antioxidant, anti-inflammatory and immunomodulating properties that may prove beneficial in modulating the excessive inflammatory activation during coronavirus disease (COVID-19) [3]. Furthermore, NAC has been extensively used as a mucolytic agent to improve airway clearance in chronic respiratory diseases.

During the COVID-19 pandemic research hypotheses on the role of NAC have been formulated and randomised control trials (RCT) are ongoing; however, to date only a few case reports have been published [3, 4].

The aim of our study is to evaluate the impact of NAC administered during hospitalisation for SARS-CoV-2 pneumonia on short-term and long-term outcomes. As short-term outcomes we considered in-hospital mortality, intensive care unit (ICU) admission, length of ICU stay and length of hospital stay (LOS) in patients discharged alive; as long-term outcomes we included diffusion capacity of the lung for carbon monoxide (D_LCO) impairment, chest radiography alterations, reduced distance walked in the 6-min walk test (6MWT) and dyspnoea score (modified Medical Research Council (mMRC) scale) at 6 months follow-up in a subset of patients included in a follow-up study. Furthermore, we will also evaluate the impact of NAC on the development of atelectasis during hospitalisation, a possible complication of SARS-CoV-2 pneumonia.

We performed a retrospective monocentric study on 1083 consecutive adult patients hospitalised for SARS-CoV-2 pneumonia at the San Gerardo Hospital, Monza, Italy, between February 2020 and April 2021. Given that the aim was to evaluate the impact of a least 5 days of NAC administration, patients were excluded if they died or were discharged within 5 days from admission (n=177) to avoid immortal time bias. NAC was introduced, as per institutional protocol, on admission and administered at a dosage of 300 mg intravenously three times daily, switched to 600 mg per os twice daily once the patient reached clinical stability and continued until discharge. The study (STORM) was approved by the national Institutional Review Board (Spallanzani Hospital), and registered at ClinicalTrials.gov with identifier NCT04424992.

As part of a multi-centre prospective study to evaluate pulmonary sequelae caused by SARS-CoV-2 pneumonia [5] (ClinicalTrials.gov identifier: NCT04435327), we also had available follow-up data on 102 patients from the original cohort alive at discharge. The follow-up consisted of a pneumological visit at 6 months including complete pulmonary function tests and D_LCO, 6MWT, mMRC scale and chest radiography.

A propensity score method was used to evaluate the impact of NAC on outcomes in the full cohort adjusting for potential confounders. We created a pseudo-population by weighting our cohort by the inverse of stabilised inverse probability of treatment weights (IPTW) computed by a multivariable logistic model on the propensity of NAC commencement with the following covariates: gender, age, days from symptoms onset to hospital admission, period of diagnosis (before/after 1 July 2020), chest radiograph (bilateral, unilateral or absence of pulmonary involvement) and ventilation support on admission, direct admission to the ICU, comorbidities (cardiovascular and cerebrovascular diseases, asthma, COPD, other pulmonary diseases, diabetes, connective tissue diseases, chronic neurological disorders, dementia and anaemia), cough and dyspnoea as symptoms and concomitant commencement of systemic steroids or remdesivir. We checked balancing among the two treatment groups after weighting by standardised mean difference and compared them by weighted two-sample rank tests [6] and a weighted logistic model. A similar approach was adopted with the subsample of subjects with a follow-up visit at 6 months.

906 patients (601 (66%) males, median age 64 years, interquartile range (IQR) 55–75 years) were included in the study, 585 (64%) received at least 5 days of NAC and 321 (36%) did not receive NAC or received less than 5 days of therapy (n=27). Demographic and clinical characteristics of the study population are summarised in table 1. Patient's characteristics were well balanced in the two groups (NAC versus others) after weighting with standardised mean difference always lower than 0.1.

In regards to the main outcomes, 133 patients died during hospital stay (91 in the NAC group and 42 in the control group), odds ratio (OR) of mortality from the IPTW weighted logistic regression was 1.22 (95% CI 0.83–1.81, p=0.3) for the NAC group versus the control group. 154 patients were admitted to the ICU with a median (IQR) of 2 (0–5) days after hospital admission and with a median (IQR) time spent in the ICU of 14 (8–24) days. After adjusting by the IPTW, no differences were observed between the two groups with regards to ICU admission and length of ICU stay. LOS in patients discharged at home was lower in NAC patients (weighted median (IQR) 15 (10–24) days) compared to those who did not receive NAC (weighted median (IQR) 17 (12–30) days, p=0.013) (table 1). Occurrence of atelectasis was not different among the two groups (2% in both groups, p=0.913) (table 1).

When considering the 102 patients (78% males, median (IQR) age 59 (53–63) years) who were followed up at 6 months with a pneumological visit, no differences were observed in regards to D_LCO impairment (weighted percentage 24% versus 19%, OR 1.35, 95% CI 0.50–3.92), chest radiograph abnormalities (weighted percentage 18% versus 14%, OR 1.04, 95% CI 0.88–1.22) and distance walked at 6MWT (weighted median (IQR) 482 (424–540) m versus 480 (432–541) m, p=0.909) between patients who received NAC (n=64) and those who did not (n=38), respectively. mMRC was not statistically different between the two groups (p=0.281), although the weighted percentage of mMRC score ≥2 (walks slower than people of the same age because of dyspnoea) was 11% in patients who received NAC and 28% in others.

Despite the preliminary evidence of a few studies, to date, there are no definitive data on the efficacy of NAC in preventing short- and long-term negative outcomes in patients with SARS-CoV-2 pneumonia [4, 7, 8]. Indeed, our results nicely fit with those of the only double-blind RCT available until now that showed no benefit of high dose NAC administration for 20 h in the emergency department on the development of severe acute respiratory failure requiring mechanical ventilation (MV) [7]. Nevertheless, LOS and long-term outcomes were not evaluated in this RCT. Similarly, a pilot study by Taher et al. [8] did not observe any benefit from NAC administered at a dose of 40 mg·kg⁻¹·day⁻¹ intravenously for 3 days in mild-to-moderate COVID-19-associated ARDS on long-term (overall mortality over 28 days) and short-term outcomes (including the proportion of patients requiring MV and changes in ARDS-severity 48 and 96 h after intervention) [8].

Among the main strengths of our study, we acknowledge the inclusion of consecutive patients from a tertiary care centre with a standardised protocol to manage COVID-19. This allowed patients to receive homogeneous treatment. The standardised protocol followed the indications from the evidence-based medicine: from March to July 2020 hydroxychloroquine, prophylactic heparin and, in cases of oxygen supplementation requirement, remdesivir were administered. Use of hydroxychloroquine was then discontinued from May 2020, due to the lack of efficacy in COVID-19 patients [9]. During the subsequent months (August 2020–April 2021) and after publication of the RECOVERY trial's preliminary data [10], corticosteroid therapy was administered to all patients requiring oxygen supplementation.

Among the study limitations the following must be acknowledged: the reasons that led physicians to administer or not administer NAC were not clear, leading to possible biases despite propensity matching.

In conclusion, our study does not suggest an impact of NAC on short- and long-term outcomes including in-hospital mortality, ICU admission, D_LCO impairment and chest radiography alterations at 6-month follow-up. Patients receiving NAC during hospitalisation for SARS-CoV-2 pneumonia presented a shorter LOS in comparison to those who did not receive NAC. However, results of the ongoing RCTs may shed further light on the role of NAC as an add-on therapy to the standard treatment for SARS-CoV-2 pneumonia.